Structural characteristics determine productivity in complex cocoa agroforestry systems

In order to cope with current challenges facing world cocoa production and the obvious lack of sustainability of the intensive model proposed to farmers, more ecologically efficient cocoa cropping systems must be developed, based in particular on a higher cultivated biodiversity level. The performances of cocoa multispecies systems, which involve multiple and hard to quantify interactions, are, however, more complicated to assess than that of monospecies systems. Despite this hurdle, we carried out a study in 48 cocoa agroforests located in three zones in central Cameroon where we conducted an analysis of cocoa yield components and agroforestry system structural characteristics that are likely responsible for observed yield variations. For the first time, we adapted the regional agronomic diagnosis method to demonstrate that the basal area per cocoa tree (mean 61.6 cm(2)) and the unproductive adult cocoa tree rate (mean 21%) are key factors when assessing the productive performance of the surveyed systems whose average cocoa yield was 737 kg ha(-1). From a methodological standpoint, the assessment approach we set up succeeded to overcome the specific obstacles linked with the features of agroforestry systems, especially their complexity (number of species and heterogeneity), by (i) determining relevant indicators and easily measurable variables, (ii) considering the associated tree communities as an environmental component, and (iii) analyzing interactions between cocoa stands and associated tree communities. From an operational standpoint, we showed that farmers can intervene on the structural characteristics of their cocoa agroforests to improve cocoa yields, in particular by eliminating unproductive cocoa trees whose basal area is less than 19 cm2 to enable the other ones to grow.